Pith. sign in

REVIEW 2 major objections 5 minor 81 references

CAMB v2's new curved-space Bessel method delivers per-mille lensed-CMB and matter-power accuracy at a fraction of the previous non-flat cost.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-02 00:51 UTC pith:HW3VAB5X

load-bearing objection A thorough, honest code-update report: the flat-Bessel Olver map is a real algorithmic improvement, and the 10^-3 convergence claim is well supported for its stated scope, though it rests on internal references rather than absolute external benchmarks. the 2 major comments →

arxiv 2607.14854 v1 pith:HW3VAB5X submitted 2026-07-16 astro-ph.CO

CAMB v2: cosmological power spectra for high-precision surveys

Antony Lewis This is my paper
classification astro-ph.CO
keywords hyperspherical Bessel functionsOlver asymptoticsline-of-sight integrationlensed CMB spectramatter power spectrumnon-flat cosmologynumerical convergenceEinstein-Boltzmann code
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper is trying to establish that the CAMB code's default settings now produce theoretical power spectra accurate enough for next-generation CMB and large-scale-structure surveys — specifically lensed CMB spectra and quasi-linear matter power converged to 0.1% pointwise on the scales that dominate the constraining power, with typical errors well below that. Its central technical claim is that curved-space line-of-sight integration can be made both fast and accurate by mapping the hyperspherical Bessel functions onto ordinary flat-space spherical Bessel functions through a leading-order Olver action construction. The paper further claims this construction is exact in the flat limit, smooth through the turning point, reduces near flatness to a simple rescaling, and removes most of the historic runtime penalty of non-flat models (about a factor 17). If true, cosmologists analyzing the next round of data can run curved and flat models at nearly the same cost while trusting the spectra at the per-mille level.

Core claim

On the paper's own terms, the discovery is a new approximation for the hyperspherical Bessel functions that appear in curved-space line-of-sight integration. Instead of matching the curved radial equation to an Airy function near its turning point, the construction matches its action integral to that of the flat spherical Bessel equation, giving a coordinate map z(χ) and an amplitude A(χ) such that the curved reduced radial function is approximated by A(χ) z(χ) j_l(ν z(χ)). The dropped term is a Schwarzian residual that vanishes exactly for flat space and is suppressed by 1/(30 ℓ^2 α^4) in the near-flat oscillatory regime. The paper argues that with this map, plus Numerov stepping seeded by

What carries the argument

The Olver action map: equality of action integrals measured from the turning point, Q_K(χ)=I_0(α z), with amplitude A=(z')^{-1/2} from the Liouville transformation, converts the curved radial equation u'' = (l(l+1)/S_K^2 - ν^2) u into the flat spherical Bessel equation plus a small Schwarzian residual. It carries the curved phase, node structure, and tunnelling onto ordinary j_l, so a fast flat Bessel routine suffices; near-flat cases reduce to a shifted argument ν_eff χ and amplitude √F0.

Load-bearing premise

The claim that defaults are 10^-3 converged rests on the assumption that the internally boosted reference runs used for comparison are close to the exact result across the whole parameter range; if that reference is itself biased in some unprobed region, the measured convergence overstates true accuracy.

What would settle it

Take a configuration outside the published validation grid, e.g. Ω_K=-0.005 with three nearly degenerate massive neutrinos and ℓ_max=6000, run the default and the strict-reference preset, and compare pointwise against the paper's own recurrence-based Bessel reference; if any interior lensed Cℓ (600≤ℓ≤3500) or quasi-linear matter-power bin (5e-3 < k h < 2) differs by more than 1e-3, or if any peak-normalized Bessel error exceeds 2e-4, the central accuracy claim fails. An independent check: tabulate the hyperspherical Bessel function by direct high-precision numerical integration of the ODE for

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Default CAMB runs can be used directly for survey-grade likelihoods without accuracy boosts: lensed TT/EE/TE and matter power meet 10^-3 tolerances on the ranges that dominate parameter constraints.
  • Non-flat models cost only modestly more than flat ones (geometric-mean CPU 5.8 s vs 4.2 s at ℓ_max=4000), making curvature scans practical.
  • The same line-of-sight machinery carries over to cross-correlations, number counts, and lensing observables, not just the two headline outputs.
  • A coherent error at the tolerance boundary would shift the survey-grade Gaussian likelihood proxy by Δχ²≈34, but measured grid values are Δχ²≈0.4, so the actual default error is well inside the envelope.
  • The recalibrated fast recombination model removes a coherent high-ℓ lensed-EE residual, with the residual recombination fit below the 10^-3 tolerance.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the action-map construction is as accurate as claimed, the technique should transfer to other curved-space transfer problems, including tensor modes and line-intensity projections, where hyperspherical Bessel evaluations are a bottleneck.
  • Because the accuracy gates are empirically calibrated on a finite grid, the strongest validation would come from an independent high-precision tabulation, such as arbitrary-precision recurrence, evaluated at random points in (K, l, ν, χ) space not on the calibration grid.
  • The paper's discussion of homogeneous reionization heating implies that the nominally 'linear' matter power at k ≳ several Mpc^-1 is convention-dependent, so survey forecasts should treat the low-redshift linear spectrum as an effective quantity rather than a unique physical prediction.
  • The documented use of AI agents for algorithm development suggests a workflow for other precision codes, but the reliability of the physics ultimately rests on the deterministic validation harness, not the development method.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper describes CAMB v2, a major update of the public Einstein-Boltzmann code. Its central technical development is a new treatment of hyperspherical Bessel functions for non-flat line-of-sight integration: a leading-order Olver action map that matches the curved radial equation to the flat spherical Bessel equation, giving an approximation that is exact in the flat limit, smooth through the turning point, and reduces near flatness to a shifted-argument flat Bessel evaluation. The paper also reports updated integrators, a recalibrated RECFAST recombination model with a helium-rate correction fit to CosmoRec, new massive-neutrino background and momentum-sampling fits, a stabilized PPF dark-energy evolution, and automatic lensing-accuracy controls. The headline accuracy claim is that default settings meet conservative 10^-3 pointwise convergence targets, relative to boosted references, over the interior lensed-CMB multipole range 600 < l < 3500 and the quasi-linear matter-power range, with typical errors substantially smaller. A deterministic validation harness is run over more than a hundred models, with per-output acceptance tolerances and a likelihood-scale proxy. A like-for-like CLASS comparison is provided for flat LCDM, and timing results show a factor-of-17 reduction in non-flat runtime versus v1.6.6.

Significance. If the accuracy and performance claims hold, CAMB v2 is a valuable community resource for Stage-IV CMB and LSS surveys. The Olver flat-Bessel construction is elegant, parameter-free at leading order, and physically well motivated; the paper's emphasis on custom-precision, gated numerical methods and on deterministic, reproducible validation is a strength. The validation harness, explicit per-output tolerance tables, honest discussion of the limits of boosted-reference convergence in Appendix C, and the clear treatment of the reionization-heating convention dependence in Appendix B are all exemplary. The transparency about the LLM/AI-assisted development process is also commendable. The main caveat is that the non-flat Bessel accuracy chain is validated against the code's own recurrence-based reference, with no fully independent numerical anchor for the non-flat case; this is the load-bearing issue discussed below.

major comments (2)
  1. [Sec. IV G / App. A 4 / Fig. 4] The non-flat Bessel validation is self-referential: the raw-Olver approximation and the empirical gates of Eqs. (A16)-(A19) are calibrated against phi_recurs, which is also the final fallback inside the same phi_olver dispatcher. The only external comparison reported for phi_recurs is a structural agreement with CLASS v3.3.4 (same seeds, recurrences, continued-fraction starts), not an independent numerical accuracy test. Because the default-vs-boosted C_l comparisons in Sec. IX use the same Bessel machinery, a shared numerical bias in phi_recurs would evade all reported tests and could push the accepted raw-Olver errors above the claimed 1e-4 envelope in unsampled corners (e.g., closed d=nu-l between about 8 and 112, or the open low-nu/l tail). I request an independent check of phi_recurs itself, for example high-precision direct integration of Eq. (4.2) or arbitrary-precision Gegenbauer
  2. [Sec. IX / App. C] The claim that stricter references demonstrate convergence is weaker for non-flat models than for flat ones. The 'strict-reference' preset increases AccuracyBoost, lSampleBoost, lAccuracyBoost, and IntTolBoost, but it does not change the Bessel evaluation path; it therefore cannot detect a common-mode error in phi_recurs or in the Olver map that passes the internal gates. Appendix C's admission that 'a single doubling does not by itself prove absolute convergence' is appropriate, and the stability under further tightening is reassuring, but for the non-flat case the entire chain of references remains internal. Please either add an independent non-flat Bessel reference as described above, or clearly delimit the non-flat accuracy claim to what the current internal-convergence argument can support.
minor comments (5)
  1. [Sec. A 3, Eq. (A16)] The gating inequality appears to contain a typo: 'h ν l > 2.5' should presumably read 'ν/l > 2.5', consistent with the following condition 'l≥50 and ν/l > 1' and with the definition of α_g.
  2. [Fig. 4 caption] The caption introduces 'd 8/(4π)' and 'd≃112' guides; the text explains the d≃112 origin, but the notation 'd 8/(4π)' is unclear and should be defined or rephrased.
  3. [Sec. V B] The values of the fitted F_He constants and RECFAST parameters in footnote 4 are given to many significant figures; given that the factor is cosmology-independent and fitted to one reference, the implied precision may be overstated. It would be helpful to state the expected sensitivity of these constants to the choice of CosmoRec settings.
  4. [Sec. XI / Table I] The timing comparison for non-flat models excludes CLASS hyperspherical-Bessel table precomputation. Since the paper emphasizes like-for-like comparisons, this should be stated more prominently in the table caption or the main text.
  5. [General] The phrase 'pointwise convergence targets' could be misunderstood: the validation compares finite sampled grids and reports maximum absolute fractional differences over ranges, not mathematical pointwise convergence. Suggest using 'maximum sampled difference' or defining the term when first used.

Circularity Check

0 steps flagged

No significant circularity: the Olver map is a parameter-free derivation; fitted accuracy gates are auxiliary and clearly calibrated, and the main validation caveat is an acknowledged limitation rather than a circular reduction.

full rationale

The central Bessel construction (Secs. IV B–IV C, App. A 1) is a leading-order Olver map: the coordinate is fixed by equality of actions, the amplitude by the Liouville transformation, C=1 by the Wronskian convention, and the Schwarzian residual Psi is explicitly dropped (Eq. A3). It is not fitted to the target spectra; in the flat limit it reduces exactly to j_l(nu*chi). The fitted elements — dispatcher gates (A16–A19), the RECFAST helium correction (Eq. 5.1), and the neutrino-background fits — are presented as calibrations against stated references (phi_recurs, CosmoRec, Fermi–Dirac quadrature), with residuals reported as calibration checks, not as independent predictions. The non-flat validation uses phi_recurs, a stable recurrence with exact seeds (A25–A27), as both the reference and the final fallback; this creates a shared-code validation caveat for non-flat models, but it is not a circular reduction: no equation in the derivation is defined in terms of the quantity it claims to predict, and no fitted parameter is renamed as a prediction. The paper itself concedes in App. C that a single doubling 'measures the distance to a more converged run, not to the exact result' — an honest limitation and a correctness risk, not circularity. External CLASS comparisons, the exact flat-limit identity, and the analytic small-chi consistency check (A15) provide independent support. Overall, the central derivation is self-contained; only auxiliary calibrations are fitted, so the appropriate finding is no significant circularity.

Axiom & Free-Parameter Ledger

8 free parameters · 6 axioms · 0 invented entities

The central Bessel scheme introduces no free parameters fitted to the target spectra; its gates are calibration constants. Auxiliary physical sub-models (recombination, neutrino fits, PPF cap) carry fitted or hand-chosen numbers, all disclosed in the text. No new physical entities are postulated.

free parameters (8)
  • phi_olver dispatcher gate constants (alpha_open, endpoint metrics, small-chi thresholds) = alpha_open=0.095/0.12(500/l)^p; p=0.70/0.14; gates 3e-3, 7e-3, 5e-2
    Empirical calibration parameters chosen so raw-Olver peak-normalized Bessel error stays near 1e-4 on a validation grid; they determine which approximation branch is used, so the central accuracy claim depends on them.
  • Accepted raw-Olver error envelope (up to 2e-4 in a small number of cases) = 1e-4 nominal, 2e-4 accepted
    The decision to accept cases between 1e-4 and 2e-4 is a measured accuracy/speed tradeoff chosen by the authors, not derived from first principles.
  • F_He helium-recombination correction constants = (a0,a1,a2,d2,d4)=(0.0781,0.899,-3.06,3.75,18.1); (z0,w)=(1968,774)
    Fitted to the total x_e(z) residual relative to CosmoRec over 1500<z_scale<3000; cosmology-independent by construction.
  • RECFAST refit parameters = RECFAST_fudge=1.125, RECFAST_fudge_He=0.84724, hydrogen Gaussian params (-0.1395,0.07299,7.281,6.767,0.1639,0.2786)
    Jointly refitted to CosmoRec at fixed RECFAST_fudge; affects the lensed EE damping tail at high ell.
  • Massive-neutrino background fit coefficients = Smooth fits, max relative error <1e-4, valid ranges 0.42<a m_nu<70 and 0.3<a m_nu<70
    Free-node fits for rho_nu, P_nu and D_nu against direct Fermi-Dirac quadrature; replace the old precomputed table.
  • PPF relaxation-rate cap beta_star = 30 (default; 10-300 give similar results)
    Chosen by hand to remove high-k/H stiffness; verified numerically against the unregularized equation at tight tolerance, but a regularization constant rather than a derived parameter.
  • Automatic lensing k_max formula = max(4,(ell_max-1500)/500)
    Heuristic cutoff formula that sets the lensing source wavenumber range; affects high-ell accuracy and runtime.
  • Free-node 4-point and 5-point massive-neutrino momentum rules = Coefficients not reported in the text
    Least-squares/nodal fits for CAMB's momentum kernels, chosen to improve velocity-dependent integrals as neutrinos become non-relativistic.
axioms (6)
  • standard math Olver uniform asymptotic / Liouville-Green comparison is valid at leading order; the dropped Schwarzian residual Psi is negligible on the gated domain.
    Invoked in Sec. IV B/C and Eq. (A3). The paper relies on empirical gates rather than a rigorous a priori bound on Psi.
  • domain assumption phi_recurs (standard three-term recurrence with exact l=0,1 seeds, Miller/continued-fraction starts) is an exact reference for Bessel evaluation errors.
    Sec. IV G and A4; it is the benchmark behind the gate calibration and Fig. 4, and is cross-checked against CLASS v3.3.4.
  • domain assumption Doubled (and optionally tripled) accuracy-boost runs in check_accuracy approximate the converged solution.
    Sec. IX and App. C; the paper explicitly states a single doubling measures distance to a more converged run, not exactness, and justifies this with stability under stricter references.
  • domain assumption CosmoRec at the stated settings is a more complete recombination reference than RECFAST.
    Sec. V B; used for RECFAST recalibration; residual differences from HyRec-2 show helium radiative-transfer modelling is reference-dependent.
  • domain assumption Linear perturbation theory with adiabatic initial conditions in the zero-acceleration frame correctly describes CMB and matter spectra.
    Sec. III; standard physics restated in the paper, not re-derived.
  • ad hoc to paper PPF relaxation-rate cap at beta_star=30 leaves observables unchanged.
    Eq. (7.5); verified numerically against the unregularized equation at tight tolerance for a set of dark-energy models, but it is a chosen regularization, not derived.

pith-pipeline@v1.3.0-alltime-deepseek · 34314 in / 14807 out tokens · 133076 ms · 2026-08-02T00:51:43.537048+00:00 · methodology

0 comments
read the original abstract

Upcoming cosmic microwave background (CMB) and large-scale-structure surveys require theoretical power spectra with numerical errors well below their observational uncertainties over the scales that carry most of the constraining power. We describe a substantial update to CAMB designed to provide fast, high-precision predictions for two key outputs: the lensed CMB and matter power spectra. The central development is a new treatment of the hyperspherical Bessel functions used for line-of-sight integration in non-flat cosmologies. A leading-order Olver construction maps the curved radial equation onto the flat spherical Bessel equation by matching their actions through the turning point. The resulting approximation is exact in the flat limit, remains smooth through the turning point, and reduces near flatness to a simple rescaling of the flat Bessel argument and amplitude. We also describe updated integrators, a recalibrated fast recombination model, stabilized parameterized post-Friedmann dark-energy evolution, and improvements to CMB lensing accuracy. Numerical convergence is assessed by comparing unboosted default results against more converged calculations. The defaults meet conservative $10^{-3}$ pointwise convergence targets over the main lensed-CMB and quasi-linear matter-power ranges relevant for future surveys. Errors measured in typical runs are substantially smaller than this. We also describe the convention dependence of the nominally linear matter power spectrum when a homogeneous calculation attempts to represent the effects of reionization heating. Essentially all of the new algorithms and code were developed with LLMs or AI agents under human supervision.

Figures

Figures reproduced from arXiv: 2607.14854 by Antony Lewis.

Figure 1
Figure 1. Figure 1: shows the resulting ionization-history resid￾uals. The upper panel shows the reference ionization history, the visibility density per unit redshift, and the positive contribution per unit redshift to the photon dif￾fusion damping scale (Silk damping) [44]. The visibility illustrates that 800 < z < 1600 contains the main last￾scattering event, while the damping kernel shows why earlier helium recombination … view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Fractional differences in the lensed [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Lensed CMB numerical validation for a flat ΛCDM model with [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Hyperspherical-Bessel pointwise-evaluation domains in the current dispatcher. Each dot is one sampled call ( [PITH_FULL_IMAGE:figures/full_fig_p020_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: illustrates the scale of the effect in a sim￾ple ΛCDM example. The linear calculation shows the expected suppression of small-scale baryon clustering from the larger post-reionization sound speed, which per￾sists to low redshift. This suppression is small on large scales, around a percent at k ∼ 1 Mpc−1 at z = 0, but grows rapidly towards smaller scales and is notionally very large, reaching several tens o… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

81 extracted references · 59 linked inside Pith

  1. [1]

    For evaluating the curved action ana- lytically it is useful to definex K =αS K(χ)

    Action integrals, the inverse map, and the leading error Withf K(χ) =α 2 −S −2 K (χ), the curved action mea- sured from the turning point is QK(χ) =    Z χt χ p −fK ds, χ < χt, Z χ χt p fK ds, χ > χ t, (A1) withQ K(χt) = 0. For evaluating the curved action ana- lytically it is useful to definex K =αS K(χ). The corre- sponding dimensionless flat var...

  2. [2]

    (4.18), the constant-map (shifted-ν) limit

    Small-χexpansion Writingz=χF,a=Kχ 2 andh=K/α 2, the differ- entiated map (dz/dχ) 2(α2 −z −2) =α 2 −S −2 K has the cubic-order solution F= 1−h 1 6 + 4a+ 13h 360 + 48a2 + 148ha+ 737h 2 45360 , (A6) D≡ dz dχ = 1 −h 1 6 + 12a+ 13h 360 + 240a2 + 444ha+ 737h 2 45360 , (A7) giving the local approximation ϕν l ≃(χF/S K)D −1/2jl(νχF).(A8) Settinga= 0 recoversF 0 =...

  3. [3]

    Forl≤2 it uses phi_recurs

    Accuracy gates For a pointwise call, the dispatching routine phi_olverdecides as follows. Forl≤2 it uses phi_recurs. Forl >2 it first tries the small-χmap if h ν l >2.5 or l≥50 and ν l >1 i , l2χ7 ν <5×10 −2, (A16) usingν/landl 2 only as cheap high-lgate proxies; the map itself uses the more accurateˆℓ= p l(l+ 1) curvature scale. If that branch is not use...

  4. [4]

    The resulting parity factor is the closed-space sign applied to allK= +1 evaluations, including the Olver map of Eq

    Recurrence and the CLASS comparison For closed models (K= +1) anyχis first folded into 0≤χ≤π/2 using the parity relations ϕν l (2π−χ) = (−1) l ϕν l (χ),(A23) ϕν l (π−χ) = (−1) ν−l−1 ϕν l (χ).(A24) which follow immediately from the Gegenbauer form be- low. The resulting parity factor is the closed-space sign applied to allK= +1 evaluations, including the O...

    2020

  5. [5]

    Bertschinger, COSMICS: cosmological initial conditions and microwave anisotropy codes (1995), arXiv:astro-ph/9506070

    E. Bertschinger, COSMICS: cosmological initial conditions and microwave anisotropy codes (1995), arXiv:astro-ph/9506070

    Pith/arXiv arXiv 1995

  6. [6]

    W. Hu, U. Seljak, M. J. White, and M. Zaldarriaga, Phys. Rev. D57, 3290 (1998), astro-ph/9709066

    Pith/arXiv arXiv 1998

  7. [7]

    Lewis, A

    A. Lewis, A. Challinor, and A. Lasenby, Astrophys. J. 538, 473 (2000), arXiv:astro-ph/9911177 [astro-ph]

    Pith/arXiv arXiv 2000

  8. [8]

    D. Blas, J. Lesgourgues, and T. Tram, JCAP1107, 034 (2011), arXiv:1104.2933 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  9. [9]

    Seljak and M

    U. Seljak and M. Zaldarriaga, Astrophys. J.469, 437 (1996), arXiv:astro-ph/9603033

    Pith/arXiv arXiv 1996

  10. [10]

    Ma and E

    C.-P. Ma and E. Bertschinger, Astrophys. J.455, 7 (1995), astro-ph/9506072. 24

    Pith/arXiv arXiv 1995

  11. [11]

    Lewis and A

    A. Lewis and A. Challinor, Phys. Rept.429, 1 (2006), arXiv:astro-ph/0601594 [astro-ph]

    Pith/arXiv arXiv 2006

  12. [12]

    Kosowsky, Efficient computation of hyperspherical Bessel functions (1998), astro-ph/9805173

    A. Kosowsky, Efficient computation of hyperspherical Bessel functions (1998), astro-ph/9805173

    Pith/arXiv arXiv 1998

  13. [13]

    Zaldarriaga, U

    M. Zaldarriaga, U. Seljak, and E. Bertschinger, Astro- phys. J.494, 491 (1998), arXiv:astro-ph/9704265 [astro- ph]

    Pith/arXiv arXiv 1998

  14. [14]

    Challinor, Phys

    A. Challinor, Phys. Rev. D62, 043004 (2000), arXiv:astro-ph/9911481 [astro-ph]

    Pith/arXiv arXiv 2000

  15. [15]

    Howlett, A

    C. Howlett, A. Lewis, A. Hall, and A. Challinor, JCAP 1204, 027 (2012), arXiv:1201.3654 [astro-ph.CO]

    Pith/arXiv arXiv 2012

  16. [16]

    Adeet al.(Simons Observatory), JCAP02, 056, arXiv:1808.07445 [astro-ph.CO]

    P. Adeet al.(Simons Observatory), JCAP02, 056, arXiv:1808.07445 [astro-ph.CO]

    Pith/arXiv arXiv

  17. [17]

    Bucher, K

    M. Bucher, K. Moodley, and N. Turok, Phys. Rev. D62, 083508 (2000), astro-ph/9904231

    Pith/arXiv arXiv 2000

  18. [18]

    Cyr-Racine and K

    F.-Y. Cyr-Racine and K. Sigurdson, Phys. Rev. D83, 103521 (2011), arXiv:1012.0569 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  19. [19]

    Tram, Commun

    T. Tram, Commun. Comput. Phys.22, 852 (2017), arXiv:1311.0839 [astro-ph.CO]

    Pith/arXiv arXiv 2017

  20. [20]

    Lesgourgues and T

    J. Lesgourgues and T. Tram, JCAP09, 032, arXiv:1312.2697 [astro-ph.CO]

    Pith/arXiv arXiv

  21. [21]

    Lewis, Phys

    A. Lewis, Phys. Rev. D70, 043011 (2004), astro- ph/0406096

    arXiv 2004

  22. [22]

    Lewis,https://cosmologist.info/notes/CAMB.pdf

    A. Lewis,https://cosmologist.info/notes/CAMB.pdf

  23. [23]

    Lewis (2026),https://github.com/cmbant/ CAMB/blob/master/docs/changelog/flat_bessel_ approximations.pdf

    A. Lewis (2026),https://github.com/cmbant/ CAMB/blob/master/docs/changelog/flat_bessel_ approximations.pdf

    2026

  24. [24]

    B. V. Numerov, Mon. Not. Roy. Astron. Soc.84, 592 (1924)

    1924

  25. [25]

    J. R. Dormand and P. J. Prince, J. Comput. Appl. Math. 6, 19 (1980)

    1980

  26. [26]

    L. F. Abbott and R. K. Schaefer, Astrophys. J.308, 546 (1986)

    1986

  27. [27]

    F. W. J. Olver, Philos. Trans. Roy. Soc. Lond. A250, 479 (1958)

    1958

  28. [28]

    F. W. J. Olver,Asymptotics and Special Functions(Aca- demic Press, New York, 1974)

    1974

  29. [29]

    Pitrou, A

    C. Pitrou, A. Coc, J.-P. Uzan, and E. Vangioni, Phys. Rept.754, 1 (2018), arXiv:1801.08023 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  30. [30]

    Trotta and S

    R. Trotta and S. H. Hansen, Phys. Rev. D69, 023509 (2004), astro-ph/0306588

    Pith/arXiv arXiv 2004

  31. [31]

    B. V. Numerov, Astron. Nachr.230, 359 (1927)

    1927

  32. [32]

    Hamann, J

    J. Hamann, J. Lesgourgues, and G. Mangano, JCAP 0803, 004 (2008), arXiv:0712.2826 [astro-ph]

    Pith/arXiv arXiv 2008

  33. [33]

    Pisanti, A

    O. Pisanti, A. Cirillo, S. Esposito, F. Iocco, G. Mangano, G. Miele, and P. D. Serpico, Comput. Phys. Commun. 178, 956 (2008), arXiv:0705.0290 [astro-ph]

    Pith/arXiv arXiv 2008

  34. [34]

    Consiglio, P

    R. Consiglio, P. F. de Salas, G. Mangano, G. Miele, S. Pastor, and O. Pisanti, Comput. Phys. Commun.233, 237 (2018), arXiv:1712.04378 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  35. [35]

    Chluba and R

    J. Chluba and R. M. Thomas, Mon. Not. Roy. Astron. Soc.412, 748 (2011), arXiv:1010.3631 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  36. [36]

    Ali-Haimoud and C

    Y. Ali-Haimoud and C. M. Hirata, Phys. Rev. D82, 063521 (2010), arXiv:1006.1355 [astro-ph.CO]

    Pith/arXiv arXiv 2010

  37. [37]

    Aghanimet al.(Planck), Astron

    N. Aghanimet al.(Planck), Astron. Astrophys.641, A6 (2020), arXiv:1807.06209 [astro-ph.CO]

    Pith/arXiv arXiv 2020

  38. [38]

    Pitrou, A

    C. Pitrou, A. Coc, J.-P. Uzan, and E. Vangioni, Mon. Not. Roy. Astron. Soc.502, 2474 (2021), arXiv:2011.11320 [astro-ph.CO]

    Pith/arXiv arXiv 2021

  39. [39]

    P. J. E. Peebles, Astrophys. J.153, 1 (1968)

    1968

  40. [40]

    Y. B. Zeldovich, V. G. Kurt, and R. A. Sunyaev, Zh. Eksp. Teor. Fiz.55, 278 (1968), [Sov. Phys. JETP28, 146 (1969)]

    1968

  41. [41]

    Seager, D

    S. Seager, D. D. Sasselov, and D. Scott, Astrophys. J. Suppl.128, 407 (2000), astro-ph/9912182

    Pith/arXiv arXiv 2000

  42. [42]

    W. Y. Wong, A. Moss, and D. Scott, Mon. Not. Roy. Astron. Soc.386, 1023 (2008), arXiv:0711.1357 [astro- ph]

    Pith/arXiv arXiv 2008

  43. [43]

    Ali-Haimoud and C

    Y. Ali-Haimoud and C. M. Hirata, Phys. Rev. D83, 043513 (2011), arXiv:1011.3758 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  44. [44]

    Lee and Y

    N. Lee and Y. Ali-Ha ¨ ımoud, Phys. Rev. D102, 083517 (2020), arXiv:2007.14114 [astro-ph.CO]

    Pith/arXiv arXiv 2020

  45. [45]

    Chluba, J

    J. Chluba, J. Fung, and E. R. Switzer, Mon. Not. Roy. Astron. Soc.423, 3227 (2012), arXiv:1110.0247 [astro- ph.CO]

    Pith/arXiv arXiv 2012

  46. [46]

    Seager, D

    S. Seager, D. D. Sasselov, and D. Scott, Astrophys. J. Lett.523, L1 (1999), arXiv:astro-ph/9909275 [astro-ph]

    Pith/arXiv arXiv 1999

  47. [47]

    M. J. Mortonson and W. Hu, Astrophys. J.672, 737 (2008), arXiv:0705.1132 [astro-ph]

    Pith/arXiv arXiv 2008

  48. [48]

    Lewis, Phys

    A. Lewis, Phys. Rev. D78, 023002 (2008), arXiv:0804.3865 [astro-ph]

    Pith/arXiv arXiv 2008

  49. [49]

    J. R. Shaw and J. Chluba, Mon. Not. Roy. Astron. Soc. 415, 1343 (2011), arXiv:1102.3683 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  50. [50]

    Silk, Astrophys

    J. Silk, Astrophys. J.151, 459 (1968)

    1968

  51. [51]

    H. H. Rosenbrock, The Computer Journal5, 329 (1963)

    1963

  52. [52]

    J. G. Verwer, E. J. Spee, J. G. Blom, and W. Hundsdor- fer, SIAM J. Sci. Comput.20, 1456 (1999)

    1999

  53. [53]

    Y. Qin, V. Poulin, A. Mesinger, B. Greig, S. Murray, and J. Park, Mon. Not. Roy. Astron. Soc.499, 550 (2020), arXiv:2006.16828 [astro-ph.CO]

    Pith/arXiv arXiv 2020

  54. [54]

    Adamet al.(Planck), Astron

    R. Adamet al.(Planck), Astron. Astrophys.596, A108 (2016), arXiv:1605.03507 [astro-ph.CO]

    Pith/arXiv arXiv 2016

  55. [55]

    S. E. I. Bosmanet al.(XQR-30), Mon. Not. Roy. Astron. Soc.512, 1046 (2022), arXiv:2108.03699 [astro-ph.CO]

    Pith/arXiv arXiv 2022

  56. [56]

    X. Fan, C. L. Carilli, and B. Keating, Ann. Rev. As- tron. Astrophys.44, 415 (2006), arXiv:astro-ph/0602375 [astro-ph]

    Pith/arXiv arXiv 2006

  57. [57]

    F. B. Davieset al., Astrophys. J.864, 142 (2018), arXiv:1802.06066 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  58. [58]

    C. A. Masonet al., Astrophys. J.856, 2 (2018), arXiv:1709.05356 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  59. [59]

    W. Fang, W. Hu, and A. Lewis, Phys. Rev. D78, 087303 (2008), arXiv:0808.3125 [astro-ph]

    Pith/arXiv arXiv 2008

  60. [60]

    Poulin, T

    V. Poulin, T. L. Smith, D. Grin, T. Karwal, and M. Kamionkowski, Phys. Rev. D98, 083525 (2018), arXiv:1806.10608 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  61. [61]

    Douspis, N

    M. Douspis, N. Aghanim, S. Ili´ c, and M. Langer, Astron. Astrophys.580, L4 (2015), arXiv:1509.02785 [astro- ph.CO]

    Pith/arXiv arXiv 2015

  62. [62]

    Lewis and A

    A. Lewis and A. Challinor, Phys. Rev. D66, 023531 (2002), astro-ph/0203507

    Pith/arXiv arXiv 2002

  63. [63]

    Lesgourgues and T

    J. Lesgourgues and T. Tram, JCAP09, 032, arXiv:1104.2935 [astro-ph.CO]

    Pith/arXiv arXiv

  64. [64]

    Hu, Phys

    W. Hu, Phys. Rev. D77, 103524 (2008), arXiv:0801.2433 [astro-ph]

    Pith/arXiv arXiv 2008

  65. [65]

    Pratten and A

    G. Pratten and A. Lewis, JCAP08, 047, arXiv:1605.05662 [astro-ph.CO]

    Pith/arXiv arXiv

  66. [66]

    Lewis, A

    A. Lewis, A. Challinor, and D. Hanson, JCAP1103, 018 (2011), arXiv:1101.2234 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  67. [67]

    T. L. Smith, V. Poulin, and M. A. Amin, Phys. Rev. D 101, 063523 (2020), arXiv:1908.06995 [astro-ph.CO]

    Pith/arXiv arXiv 2020

  68. [68]

    Challinor and A

    A. Challinor and A. Lewis, Phys. Rev. D71, 103010 (2005), arXiv:astro-ph/0502425 [astro-ph]

    Pith/arXiv arXiv 2005

  69. [69]

    Hadzhiyska, D

    B. Hadzhiyska, D. Spergel, and J. Dunkley, Phys. Rev. D97, 043521 (2018), arXiv:1711.03168 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  70. [70]

    perturbations

    Against direct Fermi–Dirac quadrature these fits reach a maximum relative error below 10 −4 in ¯ρν, ¯Pν andD ν, which is amply sufficient. The perturbations are the more delicate part. Massive neutrinos free-stream at a momentum-dependent veloc- ity, so their phase-space distribution is evolved as a sep- arate Boltzmann hierarchy for each sampled comoving...

    2000

  71. [71]

    Lewis, A

    A. Lewis, A. Hall, and A. Challinor, JCAP08, 023, arXiv:1706.02673 [astro-ph.CO]. 25

    Pith/arXiv arXiv

  72. [72]

    Casaset al.(Euclid), Astron

    S. Casaset al.(Euclid), Astron. Astrophys.682, A90 (2024), arXiv:2303.09451 [astro-ph.CO]

    arXiv 2024

  73. [73]

    Takahashi, M

    R. Takahashi, M. Sato, T. Nishimichi, A. Taruya, and M. Oguri, Astrophys. J.761, 152 (2012), arXiv:1208.2701 [astro-ph.CO]

    Pith/arXiv arXiv 2012

  74. [74]

    A. J. Mead, S. Brieden, T. Tr¨ oster, and C. Hey- mans, Mon. Not. Roy. Astron. Soc.502, 1401 (2021), arXiv:2009.01858 [astro-ph.CO]

    Pith/arXiv arXiv 2021

  75. [75]

    Salcido, I

    J. Salcido, I. G. McCarthy, J. Kwan, A. Upadhye, and A. S. Font, Mon. Not. Roy. Astron. Soc.523, 2247 (2023), arXiv:2305.09710 [astro-ph.CO]

    Pith/arXiv arXiv 2023

  76. [76]

    Lesgourgues, The Cosmic Linear Anisotropy Solv- ing System (CLASS) III: Comparision with CAMB for LambdaCDM (2011), arXiv:1104.2934 [astro-ph.CO]

    J. Lesgourgues, The Cosmic Linear Anisotropy Solv- ing System (CLASS) III: Comparision with CAMB for LambdaCDM (2011), arXiv:1104.2934 [astro-ph.CO]

    Pith/arXiv arXiv 2011

  77. [77]

    McCarthy, J

    F. McCarthy, J. C. Hill, and M. S. Madhavacheril, Phys. Rev. D105, 023517 (2022), arXiv:2103.05582 [astro- ph.CO]

    Pith/arXiv arXiv 2022

  78. [79]

    Bolliet, A

    B. Bolliet, A. Spurio Mancini, J. C. Hill, M. Mad- havacheril, H. T. Jense, E. Calabrese, and J. Dunk- ley, Mon. Not. Roy. Astron. Soc.531, 1351 (2024), arXiv:2303.01591 [astro-ph.CO]

    Pith/arXiv arXiv 2024

  79. [80]

    Lewis (2026),https://github.com/cmbant/CAMB/ blob/master/docs/changelog/hyperspherical_ bessels.pdf

    A. Lewis (2026),https://github.com/cmbant/CAMB/ blob/master/docs/changelog/hyperspherical_ bessels.pdf

    2026

  80. [81]

    Abitbolet al.(Simons Observatory), JCAP07, 034, arXiv:2503.00636 [astro-ph.CO]

    M. Abitbolet al.(Simons Observatory), JCAP07, 034, arXiv:2503.00636 [astro-ph.CO]

    Pith/arXiv arXiv

Showing first 80 references.